optimzed rotation or rotate evenly for adwords?

[mit74]

Does anyone use the rotate evenly for adverts on adwords? Is it more effective at finding the best performing ads?

I've always used optimize by clicks but I have always suspected this alters the neutral standing you want when testing new ads. It seems to put ads that Google thinks will score well rather than what you think may work well. With one campaign I had one advert taking 60% of the impressions (but actually a lower CTR than some below it) while another very similar one at just 2%. It wasn't really given a chance to see if the CTR would be better.

 

[Andrew Baker]

Always Rotate...

I like to keep the control of the ad delivery within my control and not automate it.

You should always be split testing your ad creatives, by utilising Rotate: Show ads more evenly you ensure all your ads receive a even delivery so you can make a calculated decision once you have enough data.

 

[mit74]

Always Rotate...

I like to keep the control of the ad delivery within my control and not automate it.

You should always be split testing your ad creatives, by utilising Rotate: Show ads more evenly you ensure all your ads receive a even delivery so you can make a calculated decision once you have enough data.

okay thanks. will trial it for a few months

 

[Andrew Baker]

okay thanks. will trial it for a few months

Good stuff, remember it will only work for you if you keep on top of the ads, split test 2 or 3 then analyse the results after you have enough data to make a statistically relevant decision.

Don't get blinded by CTR, obviously very important in terms of Quality Score and traffic but the main driver should be your conversion metrics.

 

[directmarketingadvice]

I've always used optimize by clicks but I have always suspected this alters the neutral standing you want when testing new ads.

It significantly distorts the result.

The point of split-testing is to show 2 (or more) versions to the same market over the same period - and the reason it's over the same period, rather than one after the other, is that markets behave differently at different times.

By choosing to optimise by clicks, the ratio of ad displays changes over time - it might start out even, but as time goes by, the higher ctr ad ends up being shown more frequently, and the lower ad shown less frequently.

This means it's not a split-test.

There's also the issue of the history of the existing ad (assuming an existing ad and a new test ad) being factored into when it is and isn't shown.

Having said all this, there's the question of whether you should be split-testing ads with PPC and whether it actually harms your campaigns.

I've got an ebook on PPC coming out in a month or so (or, basically, when I get a chance to write it) that might cover this question as, as far as I know, none of the gurus have talked about this publicly.

 

[Codefixer]

Having said all this, there's the question of whether you should be split-testing ads with PPC and whether it actually harms your campaigns.

Great, looking forward to reading this. I'm not 100% convinced that split testing all the time is the right thing to do.